Przegląd Socjologii Jakościowej, Tom XX, Numer 2, 2024
https://doi.org/10.18778/1733-8069.20.2.04

Kamil Pietrowiak ^*

University of Gdańsk, Poland
https://orcid.org/0000-0001-7690-7811

“Follow the Dog”: Using the Go-Along Method in Research on Training and Working with Guide Dogs for People with Visual Impairment^[1]

„Idź za psem”. Zastosowanie metody go-along w badaniach nad treningiem i pracą z psami przewodnikami dla osób z niepełnosprawnością wzroku

Introduction

Training a guide dog is a multifaceted and long-term process, involving different types of actors and depending on variables that are not always controllable. To create conditions for a safe and effective cooperation between dogs and visually impaired handlers, it is necessary to secure not only adequate funding, but also expertise and contribution of many parties – breeders, NGOs, volunteers (puppy raisers), and trainers. Another vital element is access to dogs that are healthy, non-aggressive, cooperative, motivated, and showing both initiative and obedience. Then, of course, it is essential for the guide dog to be expertly trained. Finally, the determination and adequate preparation of visually impaired handlers – and, in particular, their willingness to assume responsibility for the guide dog – are yet another key factors.

Preparing a dog to guide a person with visual impairment requires many hours and miles of walking together with the dog, both with volunteers responsible for socializing a puppy and with trainers teaching not only specific skills related to different elements of space, but also training the dog to ignore distractions and general obedience. An important step in this process is the transition from teaching single ‘tricks’ to combining them as part of longer routes, and from guiding the dog’s actions by the human to giving the initiative and agency back to the guide dog. The next step is carefully matching the dog and their prospective handler, taking into account both parties’ personality, size, and walking pace. Once a promising team is successfully matched, there begins an intensive period of learning how to communicate and interact, recognize each other’s needs and capabilities, and build trust and rapport. This process continues beyond the end of official classes led by trainers and takes place, though with varying intensity, throughout the lifetime of the rapport between the visually impaired handler and the guide dog. Thus, training a guide dog cannot be compared to programing a device, since it involves the emergence and consolidation of the dog’s habits and skills that require constant practice, care, and work on the part of the handler, at risk of their extinction should such care be missing.

In practice, guide dogs spend most of their time as regular pets – playing, going for walks, resting, sleeping – but the way they are perceived is associated with a ‘professional’ role they perform outdoors wearing a suitably marked harness. In this context, the dog’s tasks include safely guiding the person in a designated direction, avoiding all kinds of obstacles, targeting pedestrian crossings and stairs, stopping in situations of danger or a lack of passage, and pointing to elements of space in line with the handler’s command. The competences of a visually impaired handler include: giving commands that the dog can understand; orientating in space and determining the purpose of the walk; interpreting the dog’s movements and decisions; motivating, rewarding, and correcting the dog; taking care of the dog’s safety; and satisfying the dog’s needs.

Despite the dynamic nature of the training of guide dogs and their subsequent collaboration with visually impaired handlers, much of the research on this topic has been based on questionnaires and individual or group sit-down interviews (Woods, 1994; Sanders, 1999: 39–58; Modlin, 2000; Wiggett-Barnard, Steel, 2008; York,Whiteside, 2018; Lieberman et al., 2019). So far, few researchers have chosen participant and non-participant observation or walking interviews with visually impaired handlers and guide dogs as their main or complementary method of investigation (Naderi et al., 2001; Magnus, 2014; Bohan, Tuck Wah James, 2015; Due, Bierring Lange, 2018). Some studies provide no detailed description of the applied observational methods, nor a thorough analysis of data obtained in this way. In these cases, authors focus on discussing quotes from interviews or messages exchanged with participants (Sanders, 2000; Porkertová, 2020).

Recognizing this deficit, I opted for the systemic use of the go-along method in research project on the collaboration and communication between visually impaired handlers, guide dogs, and other participants in the training process (project title: “Cooperation between blind people and guide dogs. Cultural and social determinants of the interspecies relationship”). The aim of this article is to show the potential, benefits, and dilemmas related to the use of this method in this particular research project, including technical and ethical aspects that I needed to take into account in my research design and its implementation. Apart from general considerations and technicalities, I will also briefly discuss select research data generated via the go-along method. Before I discuss the project itself, I would like to discuss basic premises and ways in which the go-along method has been applied by other researchers, along with their reflections on the advantages and limitations of combining mobile observation with interviews.

The go-along method – the premises, potential, and practical application

The go-along method is usually associated with the so-called ‘spacial turn’ and ‘mobility turn’ in social sciences, and with embodied approach to social research (Sheller, Urry, 2006; Büscher, Urry, 2009; Rogowski, 2016: 123–125). The development and mutual complementarity of these paradigms has alerted researchers to the organic interdependence between people’s practices and strategies, and their environment – places, objects, and other organisms. In this approach, space is not a collection of points that exist objectively and independently, but a dynamic network of connections and relationships between different kinds of elements in constant flow and interaction. Hence, instead of asking how people operate in a predefined, physical space, researchers began to consider how people resonate with their environment as part of everyday life dynamics, movement, and action. In this sense, people are not users of passive space, but, instead, they are both co-creators of and co-created by their environment. The way humans experience and perceive space is not static and ready-made, but relational and contextual, dependent on current perspective, emotions, biological and material possibilities, limitations, and ways of moving. In short, they are predicated on all physical, cultural, and social conditions (Urry, 2000; Larsen, Urry, Axhausen, 2006; Ingold, 2011).

The go-along method constitutes an attempt to translate these theoretical premises into social research practice. Generally, this involves purposefully moving together with research participants within a specific space, combined with interviewing and participant observation. The aim is to capture – to see, hear, and feel – how research participants experience and interpret elements of a given environment; how they interact with them; how they navigate and find their way through different kinds of spaces; and how they co-create paths, intersections, points, and places (Kusenbach, 2003: 463; Bates, Rhys-Taylor, 2017: 2).

The go-along method can be seen as a more focused and structured variation of participant observation or ‘hanging out’ with key research participants (Kusenbach, 2003: 463). The main difference is that the go-along method is deliberately used in dynamic settings and focuses on the experience of specific aspects of reality, especially movement and space. Some researchers pointed to its similarity to the shadowing technique (Evans, Jones, 2011: 850; Butler, Derrett, 2014: 1–2), but the difference between these two is apparent when we consider two aspects. First, in the go-along method, the relationship between research participants and the researcher tends to be more casual, based on partnership and spending time together, whereas shadowing leaves less space for active interaction and unrestricted interviewing by the researcher, who often assumes the role of non-participant observer. Second, the shadowing technique is usually applied in the context of organizations – in particular, internal structures and processes – where it is used for observing selected individuals in their professional roles, activities, tasks, and relationships (Czarniawska, 2007; Quinlan, 2008; Sirris, Lindheim, Askeland, 2022).

The premise of the go-along method is ‘talking whilst walking’, hence it is sometimes also referred to as the ‘go-along’ or ‘walking interview’ (Anderson, 2004; Butler, Derrett, 2014; Marcotte et al., 2022). The walking interview can have a variety of forms, depending on the research setting, topic, or group, but most often it comes as either an unstructured or a semi-structured interview (Carpiano, 2009: 265; Bell, Bush, 2021: 316). From the practical point of view, the greatest advantage of this method lies in its hybridity and complementarity between multisensory participant observation with different forms of interviewing. Combining these techniques makes it possible to follow both steps and thoughts of research participants, and, depending on our research interests, to inquire and direct their attention to selected practices, experiences, emotions, and memories related to specific sites (Kusenbach, 2003; Hein, Evans, Jones, 2008; Carpiano, 2009).

Most often, researchers walk with the research participants (go-alongs), or use other modes of transport: cars, bicycles, or public transport (ride-alongs) (Spinney, 2011; Wegerif, 2019). In most cases, the researcher joins an individual; less often, the walk and interview involves a couple or a larger group (Clark, 2017; Marcotte et al., 2022). Apart from observation, conversation, and listening, researchers record walks and rides together in a variety of ways. The most basic methods involve taking notes during or after the meeting, photographing selected locations, and recording the conversation with a voice recorder (Carpiano, 2009; Adekoya, Guse, 2020). Some researchers choose to film such walk-alongs, which some believe to be the most suitable way of recording such encounters, capturing their dynamics and the environment (Pink, 2007; Parent, 2016; Vainini, Vannini, 2017). The camera can be operated by the researcher, who is simultaneously talking to the research participants, or by another person from their team (Rogowski, 2016: 132–133). In addition, some researchers, especially working in the domain of social and cultural geography, use GPS and GIS technologies to record the exact course of walk-alongs, generating various types of maps (Hein, Evans, Jones, 2008; Evans, Jones, 2011).

An important criterion in the go-along method is the choice of route. In many cases, researchers are keen to have the space and route of a walk-along determined by the study participants, which may mean moving freely through a familiar area or moving toward a specific destination. Another option is to invite research participants to walk around places that are unfamiliar to them and along routes predetermined by the researcher. Between these two possibilities, there are also intermediate solutions, combining the initiative of research participants and researchers (Evans, Jones, 2011: 850; Rogowski, 2016: 130). Importantly, the choice of sites and spaces to be explored is virtually unlimited – often it is the research participants’ own neighborhood, but it may also be other public places, green areas, homes, or institutions (Macpherson, 2016; Clark, 2017; Duedahl, Stilling Blichfeldt, 2020; Marcotte et al., 2022: 2).

Due to its versatility and flexibility, the go-along method in its different guises can be helpful when exploring a wide range of topics and is thus used by researchers from the fields of sociology, cultural anthropology, cultural geography, health studies, or disability studies. As Margarethe Kusenbach (2003), the pioneer of this approach, pointed out, this kind of dynamic co-existence can be fruitful for research on perception, spatial practices, personal biographies, social architecture, or social realms. This author’s intuition is corroborated by examples of the go-along method used within projects on, e.g., youth leisure time (Clark, 2017), the impact of different types of trauma on the experience of space (von Poser, Willamowski, 2020), pedestrian strategies and walkability assessment (Battista, Manaugh, 2017), outdoor learning and education (Lynch, Mannion, 2016), public parks and inclusive openness (Neal et al., 2015), or routes and modes of food transportation in metropolitan areas (Wegerif, 2019). In addition to its cognitive value, the go-along method can have practical or social benefits, as demonstrated in cultural animation projects (Rogowski, 2016), spatial consultation (Bergeron, Paquette, Poullaouec-Gonidec, 2014), or the empowering of groups at risk of exclusion (O’Neill, 2018).

For the purposes of this article, it is particularly relevant that the go-along method was used for research that involved people with specific motor, bodily, and perceptual capacities and limitations, stemming from their health or (dis)ability and affecting their ways of moving, traveling, and navigating space (Bartlett et al., 2023). In this context, it is worthwhile to consider research on the disabling environment and strategies used by people with disabilities to deal with the inaccessibility of public space (Blewett, Hanlon, 2016; Wästerfors, 2021), on mobile and spatial practices of wheelchair users (Parent, 2016), on factors that increase the independence of autistic people at home (Marcotte et al., 2022), or on ways in which older adults with dementia move, navigate, and interact with others in care homes providing long-term care (Adekoya, Guse, 2020). The go-along method has also been helpful in researching the ways in which people with visual impairment perceive and interpret different types of places, navigate and move through familiar and unfamiliar space, and use available information and technology (Macpherson, 2009; 2017; Due, Bierring Lange, 2018; 2019).

Like any other research method, it also has its limitations. One of the most commonly mentioned restrictions is dependence on external conditions – the weather, in particular – which may prevent or hinder walk-alongs (Garcia et al., 2012; Castrodale, 2018: 50; Thompson, Reynolds, 2019: 6). The unpredictability that is inherent in walk-alongs – which prevents replicability, standardization, and full control over the setting – may also present as a limitation (Duedahl, Stilling Blichfeldt, 2020: 456). Researchers may find it challenging to multitask – listen, talk, observe the environment, and study participants while also operating a video camera (Evans, Jones, 2011: 851; Rogowski, 2016: 132). Some complications may arise from the attitude or mindset of research participants who may prefer to be interviewed in a way that they find familiar (sitting down to have a conversation) or who may not see the point of walking interviews (Castrodale, 2018: 8; Marcotte et al., 2022: 6). Finally, visual recording of walks raises questions about the form and conditions of subsequent audiovisual material presentation. According to Phillip Vannini and April Vannini (2017), researchers often emphasize that walking interviews have an embodied and multisensory nature but fail to present video footage to illustrate and confirm that claim. Also, publishing video footage requires the research participants’ consent – pertaining to copyright and right of publicity – which makes it necessary to reflect on the legal and ethical issues resulting from such specific research design (Rogowski, 2016: 135–136).

The use of the go-along method in research on training and working with guide dogs

In this paper, I discuss the implementation of the go-along method within the research project that focuses on cooperation, communication, and rapport between guide dogs and visually impaired people, trainers, and other subjects, conducted between 2021–2024 in Poland and Slovakia. The participants in the study included NGO members in charge of managing training and handing over guide dogs to their handlers, junior and senior trainers, volunteers who raise the dogs for the first 11–13 months of their lives, and such organizations’ clients who wish to benefit from a guide dog’s assistance (hereafter: candidates), as well as people with various degrees of visual impairment who live and move about with guide dogs – from beginners to experienced handlers who have already worked with more than one or even two dogs.

In the project, I decided to use a variety of qualitative methods – IDI interviews including elements of biographical interviews, conceptual interviews and projective techniques, participant observation, and the go-along method. The participants I interviewed and accompanied during individual walking interviews had signed their written consent to participate in the study. While observing group training sessions, I asked their organizers to obtain verbal consent from the participating candidates or, wherever it was possible, I had personally asked for such consent before the session^[2]. The project’s research design has been approved by the Research Ethics Committee of the University of Gdańsk (approval no. 4/2022).

So far, I have conducted 109 IDI interviews (including 6 interviews with both the participant and their partner), lasting between 50–220 minutes (most interviews: 110–130 minutes). I interviewed 85 people, including 32 with different degrees of visual impairment. I interviewed some trainers and dog handlers on 2–3 different occasions in order to fully benefit from their experience and expertise as well as their willingness to share their stories and reflections.

Participant observation took place in a variety of settings, such as: during puppy testing, volunteer training and coaching, specialist dog training sessions, dog and trainer examinations, training and qualification tests for candidates applying for a dog, the process of matching a guide dog with a handler, the process of handing over the guide dog, dog and handler examinations, check-up meetings (6 and 12 months after the dog’s hand-over), independent work with the guide dog by a visually impaired handler, and visits to the visually impaired handlers’ homes. My participation in training-related events for dogs and dog handlers was possible with the consent given by visually impaired handlers, but also the kind permission and support of the organizations and people managing the process. In total, I spent around 780 hours on observation, including about 360 hours when I joined the study participants on various types of walks and journeys with the dog.

It this particular project, the go-along method was one of the main rather than an auxiliary way of data collection. The method involved my attentive and purposeful presence during the joint movement of the team made of a guide dog and human handler (including puppy raisers, trainers, candidates, and visually impaired handlers). As I developed the study design, the use of the go-along method seemed an obvious choice, tailored to the very nature of the collaboration between study participants and dogs, which mostly takes place outdoors, and largely on the move. I assumed that in order to develop an understanding of how they communicate and cooperate, it was necessary to walk together, follow the same routes, and participate in training sessions. In short, I needed to put on some comfortable shoes and walk as much as possible with, following, and beside people and dogs.

The people I invited to participate in the study understood the purpose of me accompanying them as they worked with the dog and did not challenge the study design. To ensure their comfort and facilitate their agency, I adapted to their decisions regarding the length of the walk-along, the specific route, and other details^[3]. In cases where either the person or the dog were having a ‘bad day’, i.e., were feeling tired, ill, or injured, as well as when weather conditions were unfavorable (heavy rain, very hot, or very cold weather), we only proceeded with a classic sit-down interview (Carpiano, 2009: 269).

In terms of time, accompanying study participants during their training sessions and walks ranged from 7 minutes for absolute novice handlers, who usually showed me their routine route near their home, to 8 hours, for full-day training sessions with candidates or individual meetings with guide dog handlers. It would be difficult to provide the exact count of either steps or hours walked-along for longer meetings, as we also had rest and meal breaks, conversations, and obedience exercises for dogs that did not involve moving about. In such situations, the go-along method was used in tandem with ethnographically informed ‘hanging-out’ (Kusenbach, 2003: 463). Additionally, some training sessions and daily walks involved the participants using public transport, where we switched from walk-along to ride-along perspective (Wegerif, 2019).

The practical implementation of the go-along method varied depending on several factors. First, walk-alongs were divided into individual and group walks. Individual walks mostly took place when I had longer meetings with trainers or visually impaired handlers – usually before or after a sit-down interview. During the walk, I followed either two steps behind the human and the dog or to their right. The choice of exact position depended on the dog’s reaction to my presence – the first option was advisable if my presence distracted the dog or if the handler needed to be more focused on ‘reading’ the dog’s movements. Walk-alongs during group walks, on the other hand, took place at various stages of the dog and handler training process: starting with training sessions and various tests for puppy raisers and younger dogs, through the spatial orientation exams for candidates and meetings aimed at matching them and dogs, to training sessions before the final hand-over of the dog. Such events involved a minimum of two people (a trainer and a puppy raiser/a candidate) and a dog, but it was not uncommon for other trainers to be present. During training sessions, I usually walked two to six meters behind the session participants.

Second, the above-mentioned situations were either initiated or triggered by me, or they happened independently of my initiative, pre-planned by research participants. The first type refers mainly to individual encounters with visually impaired handlers whom I asked for a walk-along with their guide dog, before or after a sit-down interview. On such occasions, even if participants did some shopping or allowed the dog to walk at leisure, it was not part of their daily routine and previous plans, and was considered to be part of the study. The second type mostly involved training and coaching events, which I was able to attend with the consent given by the individuals involved. During some walk-alongs with particular visually impaired dog handlers, I was also able to join them during their daily activities, not initiated by nor due to my presence. These were usually routes to or from work, shopping, a visit to a doctor or a vet, or picking up a child from the kindergarten.

Third, in terms of attention and involvement modes in such situations, my activities ranged from walking observation to walking interview. In particular, I focused on observation during training sessions with puppy raisers or candidates, when I watched their complex cooperation with dogs and trainers, and listened in to their ongoing communication. In such cases, I preferred not to interfere and only asked questions during breaks or longer stops. If there were other trainers present and walking behind the triad or the group, I talked to them or paid attention to their comments about how the particular dog and the handler were doing.

Interviewing was possible during individual walk-alongs with those dog handlers who were able to both concentrate on the route and answer my questions or comment on interesting or salient dog behavior and some aspects of space. Also, I could interview most of the participants as we followed quieter and simpler sections of specific routes. During walk-alongs, I used unstructured interviewing, pertaining to the ongoing elements of dog-human cooperation and communication, or to issues that emerged during our sit-down interviews. In practice, the focus of my attention and activities kept changing depending on the situation, the degree to which participants could focus on multiple tasks, the difficulty of the route, or the intensity of the soundscape. Interactions oscillated between my silent observation and our lively conversation, which tended to be strictly related to the research topic, though occasionally touched upon other issues.

Some parts of walk-alongs were recorded using a small sports camera mounted on a selfie stick or, in emergency situations, with a mobile phone (the total number of 25 hours of footage). All participants have given their written or verbal consent for recording to be made. The frame usually included the dog and part or all of the handler’s figure, seen from behind or from the side – depending on my position during the walk-along. During several longer walk-alongs with visually impaired handlers and their guide dogs, which became walking interview as the participants were experienced handlers, I also additionally used a voice recorder clipped to the interviewees’ clothing to guarantee good sound quality (Adekoya, Guse, 2020). When I made no video or audio recording, and when my activity was focused on observation, I also used to jot down notes on my mobile phone as we walked.

Photograph 1. A trainer and a candidate during a session – captured from the vantage point of the researcher following them at a distance of 2 meters
Source: photograph by the author.

Photograph 2. Walk-along with a visually impaired dog handler. The researcher was walking to the right of the pair, interviewing the handler
Source: photograph by the author.

Photograph 3. Walk-along with a visually impaired handler and a guide dog. The researcher was walking a few meters behind the participants
Source: photograph by the author.

There were several reasons for which I opted to use a sports camera in a specific way. First, I chose not to attach the camera to the study participants’ clothing, as this would not reflect their perspective, which was relevant not only for people with visual impairment, but also sighted research participants, who often looked at the dogs and kept eye-to-eye contact with them instead of looking straight ahead (Battista, Manaugh, 2017). Upon consideration, I decided not to ask sighted participants to place the camera on their heads, as this could have been disturbing and intrusive, and may have discouraged them from participating in our research project. Second, I did not film the participants by standing in front of them, as this could entice the dog to follow me and reduce the dog’s independence in guiding the handler (Due, Bierring Lange, 2018). Third, keeping the camera in close proximity to the participants generally allowed me to record their comments, answers to my questions, and in particular, their verbal communication with the dog. Finally, the appealing visual quality of the recorded scenes was less relevant, as I did not plan to present the footage to a wider audience.

The participants’ written consent forms included a clause on filming of their cooperation with the dog as well as its subsequent analysis and archiving by the researcher. When filming groups, I sought verbal consent from all the group members. As I turned the camera on, I usually reiterated that the footage was just for me to use and that I was not planning to publish it on the Internet – that I would watch it later to detect details I may have missed. The choice not to use video to present results was primarily ethical. I wished to avoid using my visual advantage over the visually impaired research participants or expose them to the gaze and potential assessment by the viewers. The fact of their visual impairments could make them vulnerable and, as such, required the researcher to be careful and reflexive about the overall research design (Hammer, 2019: 15–18; Surmiak, 2022). At the level of fieldwork and in specific situations, this was particularly true for novice dog handlers who were just getting to know their guide dogs. Seen from the outside, without the wider context of the training process, their behavior could be interpreted as unskillful, and the dog could be perceived as ‘poorly trained’. Therefore, in this article I do not provide links to any video material, but only selected frames, in which faces of research participants are not visible. Some footage will be presented at conferences and other scientific events, provided consent for such use of the footage is obtained from individual participants.

The recordings of training sessions with guide dogs and trainers or volunteers pertained to internal procedures and methods applied by specific organizations, and I assumed that as such, they should not be made public without these organizations’ oversight. To sum up, my rapport with research participants and the nature of our encounters were both based on trust and openness. I believed these may have been compromised should the participants know that the collected video footage would be published on-line, beyond our control, and be interpreted in a way that would not take into account the context and the rapport in question.

At the stage of organizing and analysing the collected video footage, I used MAXQDA, a software that enables collection, organizing, analysing, and visualizing data collected through qualitative and mixed methods. It makes it possible to transcribe video footage, add notes or paraphrase text at specific points in the video, encode selected excerpts, view the material in slow motion or fast-forward mode, and export clips and images from the videos (Kuckartz, Rädiker, 2019: 83–92). These functions allowed me to avoid tedious, elaborate transcription that would need to include additional description of the audiovisual context (Rogowski, 2016: 139). It was also important that this tool made it possible to incorporate video footage within the overall network of other kinds of collected research data, facilitating its use for further, enhanced analysis.

In my view, the key features were the possibility to add notes and encode selected scenes, which facilitated rapid cross-referencing of the interviewees’ stories, recorded events, relevant literature, and other collected material. Such cross-referencing allowed me to see what is heard and hear about what is seen, and supported the emergence of a multi-level, grounded knowledge of the phenomena in question.

The go-along method enabled me to collect valuable data on different stages of training and working with guide dogs. Accompanying the volunteers during training sessions and tests revealed what skills the trainers expected the dogs to have developed at different stages, as well as various levels of experience and differences in approach among the volunteers, manifesting in their verbal and non-verbal communication with the dog. Following the trainers and dogs during the course of more targeted training sessions allowed me to observe specific training techniques and methods, and to inquire about their relevance in the context of preparing the dog for guiding a person with visual impairment. Walks with dogs, trainers, and candidates revealed challenges and difficulties involved in handing over a dog to a visually impaired handler, as well as the dynamics of interaction between the different actors in such a situation. Finally, walking and moving with the guide dogs and their visually impaired handlers made me notice many details of their verbal, sensory and bodily communication, as well as the ways in which they worked together in a specific environment.

Due to the specific objective and scope of this article, in the following sections I only briefly discuss selected data generated via the go-along method, focusing exclusively on those stages of the training process and guide dog work that involved people with visual impairment. Thus, the following sections do not constitute an in-depth analysis. Their primary purpose is to show the kind of data that were observed, heard, elicited, and collected using the go-along method in this research project, and the resulting benefits in terms of understanding specific themes related to the work of guide dogs for visually impaired people.

“Follow the dog” – walking along trainers and novice guide dog handlers

The culmination of the guide dog and candidate training is the dog hand-over to the visually impaired handler, which takes an average of 80 lesson-hours, continuously or spread over a period of 2–3 weeks, first in the city where the organization is located and then at the candidate’s place of residence. The overall goal of these sessions is to learn communication and develop trust between the dog and the dog’s new human handler, so as to lay the foundation for their independent and safe mobility.

Participation in such training sessions gave me an extraordinary opportunity to observe the interactive dynamics within the trainer-dog-candidate triad, which has so far been very rarely the focus of scientific publications related to guide dogs^[4]. The hand-over stage is an extremely intense and challenging time for everyone involved in the process – for the dog, who is suddenly required to guide a stranger, whose verbal and non-verbal communication is often not clear and ‘readable’ for the dog; for the trainer, burdened with great responsibility for the safe future of the dog-human tandem; and for the visually impaired person, who – if they are novice dog handlers working with a dog for the first time – needs to find their way through a flurry of new information and tasks involved in working with the guide dog. Given the scope of this article, below I provide only some examples of how trainers imparted their knowledge on working with a guide dog, in situations I recorded using the go-along method.

For the first day or two of the hand-over, the trainers walked on the left side of the team (the dog’s side), with an extra leash attached to the dog’s collar, to enable them to intervene in emergency situations, when the dog – distracted by a new situation – would not stop at the required place. When trainers decided that the dog and the candidate were ready to walk independently, they would detach the extra leash and walk at a greater distance, either to the right or behind, making it difficult for the dog to turn around and look for contact with the previous handler.

Regardless of the hand-over stage, safety remained an absolute priority for the trainers, who constantly raised the candidates’ awareness in this matter.

[Speaking formally to the handler] Are you at the pedestrian crossing? We need to know that. Are you at the pedestrian crossing or not? She [the dog] claims she is. I say, not necessarily. You must be absolutely sure if you are at the pedestrian crossing or not. Please correct. Once again, “correct”. Do you feel the blisters [underfoot] now?

[Speaking informally to the handler] When you let the dog cross a low kerb once and then don’t let him another time, it gets all mixed up in the dog’s head. The dog doesn’t think as logically as humans do. With fixed routes, it might be OK, but if the dog does not stop when you’re doing a new one, it can be dangerous. You let the dog make a small mistake, then a bigger one, and then, finally, the dog starts to ignore it altogether and stops targeting.

[Speaking informally to the handler] I prefer to hold you a little bit because you’re so rushed [there are stairs leading down]. And she [the dog] also walks fast and then halts before the stairs. It is uncomfortable, can you feel it? We need to slightly increase the pressure on the harness a bit earlier, so you don’t lose contact with the dog.

Emphasis on safety combined with raising the candidates’ awareness that they should be constantly attentive to the dog’s actions, consistent in their messages, and aware of their location within the given area. Also, in order to be able to work effectively with the dog, visually impaired handlers were required and in practice had no other choice but to trust the dog and let the dog take some initiative and agency while guiding. For this reason, trainers consistently highlighted the need to trust the dog’s abilities, to make space for the dog to act, and to wait for the dog’s decision. During training, this was usually expressed in short sentences, such as: “Follow the dog”, “Let the dog do the work”, “Do not take over”, “She’s going well, do not push her”, “Give the command and don’t suggest anything”, “Do not turn yourself, let the dog earn the reward”. Similar cues were sometimes expressed more elaborately, eliciting the way in which the trainers thought about the relationship and interdependence between the dog and the visually impaired handler.

[Speaking informally to the handler] Remember, the dog goes first, the handle follows the dog, then follows your hand, and finally you. So, you do not move until you feel the handle moving forward.

[Speaking informally to the handler] Let me explain something to you. If you push her, she won’t go. This is your mistake, you wish to go too fast. You push her and then she would not budge, like a donkey. Level off and let her move on her own. See what you do with the handle? You’re putting pressure again. Why should she do it, when you do it for her? These details are important. You have to go back to the beginning. Because if you’re pushing the dog now, before long you’re going to have to do her work and show her where to go.

Apart from the need to give the initiative back to the dog and ‘make space’ for the guide dog, trainers continually stressed the importance of rewarding the dog for a task well done in the form of both a treat and a voice praise in order to motivate the dog to work. At this stage of training, frequent rewards and praises were also intended to make the new handler more attractive to the dog, who would associate the new person with something pleasant and, as a result, focus the dog’s attention on the handler and facilitate establishing good rapport between them.

[Speaking informally to the handler] Give her a treat in front of the escalator, so that she stops and not rush forward. Give her a treat once she’s on the stairs so that it’s nice.

[Speaking informally to the handler] She has targeted the walkway, so you should praise her. She thinks to herself: “I’ve shown something, but what’s in it for me?”. You need to give her a message. She needs to work for something.

More often than not, the strategy worked as the trainers intended. In cases where I was able to accompany a particular dog for a few days during the handover process, I observed the dog gradually shifting their attention from the trainer to the visually impaired handler, glancing behind less and less, and increasingly directing their gaze toward the face of the “new human” and their snack pouch. In addition, as days went by, the dogs became more and more responsive to the commands given by candidates, without waiting for the trainer’s confirmation.

It was equally common for trainers to describe to candidates the immediate environment along with the dog’s current behavior and reactions. This not only was due to the nature of the limitations imposed by visual impairment and the lack of the candidate’s sensory access to these signals, but also stemmed from the highly visual character of training and communication with the dog (Coren, 2000; Ittyerah, Gaunet, 2009). In this context, it was crucial for trainers to translate the visible elements of the dog’s body language into non-visual cues sensed through the leash and on the harness handle. Candidates also had to learn to think ahead, predicting and anticipating the dog’s behavior in certain situations, e.g., when hearing another dog barking or squealing nearby.

– [Speaking formally to the handler]What could we be facing here? Do you know what it could be?
– The lawn.
– Well, the smells. So, she must forget something.

[Speaking informally to the handler] There’s going to be a surprise right in front of us. A little dog. You stop, because you don’t know what is in front of you, and the dog has turned off. You stop, you reset the dog by saying the dog’s name, and the name means “look at me”. Or you say “enough”. Or alternate the two. To ask the dog to pay attention to you.

When verbal cues were not sufficient to form an understanding of the situation, the trainers tried to convey them by touching or directing the candidates’ bodies. In some cases, this proved to be the easiest way to explain certain technical points. At the same time, this kind of interaction evidenced a personal, bodily nature of the rapport between the trainer and the candidate, one that required mutual trust and awareness of each other’s boundaries.

[Speaking informally to the candidate] Also, remember not to overtake the dog now. Because, now you want to walk in this position [at the level of the dog’s front paws]. And your position while walking – sorry, I am going to pull you a bit – is here [at the level of the dog’s hind paws]. That is, while walking. Because when you stand still, it is like you did, halfway or at the shoulder. But when you’re walking, make sure that you move a little bit away. So that it’s not like she’s saying: “I don’t have the space to make this turn.”

It is difficult to imagine collecting this kind of data without the go-along method applied during training sessions with trainers, dogs, and candidates. Although the insights on the hand-over stage were also discussed in detail during individual interviews, they turned out to be more general, filtered through memory and temporal distance. The same was true for the knowledge imparted to the candidates during sit-down classes, which were more theoretical in nature. It was only by walking along the trainers, candidates, and dogs that I became aware of the complex, bodily, multisensory nature of both knowledge and skills that people with visual impairment need to acquire to work with their first guide dog. It was also through walk-alongs that I learned that working with guide dogs is predicated on binary oppositions – simultaneous control and trust, being a leader and the one who is being guided, thinking ahead and trusting the dog’s skills.

“Forward, don’t cheat” – verbal cues for guide dogs

One of the elements on which I focused during our walk-alongs with visually impaired handlers was the forms of verbal communication they used with the dog. Some participants limited themselves to uttering commands and instructions (“no!”, “don’t!”), while others proved to be more talkative, occasionally engaging in a kind of monolog/dialog with the dog. Over the course of subsequent walk-alongs and interviews, I began to discover that the frequency and manner of speaking to the dog depended very much on the person’s personality, their general attitude toward the dog, the nature of their relationship, the dog’s reaction, and the complexity of the current situation on the route.

The basic form of communication when working with a guide dog is uttering standard commands relying on the dog’s skills. The commands express the handler’s expectations as to starting, direction, pace, and stopping the movement (“forward”, “right”, “left”, “hup-up”, “steady”, “stop”, “retreat”), as well as the expectation that the dog will find and target a particular element of space (“crossing”, “walkway”, “stairs”, “up”, “down”, “door”, “bench”, “bus stop”). In one school, the trainers added the words “lead” and “show” to the above commands (e.g., “lead forward”, “show crossing”). Also, some commands refer to a specific position of the dog and are used both in working mode and at home (“sit”, “down”, “stand”, “heel”, “stay”, “come”, “place”). Some commands refer to specific skills, such as deciding independently how to avoid an unusual or unexpected obstacle (“lead yourself”) or finding and passing dropped objects (“find”/“pass”). Importantly, it is not only the content but also the manner in which the commands are given that is important for the collaboration between a dog and a human. Commands need to be uttered using appropriate intonation, firmness, separated from other words. In addition, some trainers – and by extension the visually impaired handlers they train – pronounce selected commands with a characteristic sing-song quality intended to make it easier for the dog to read the cue.

Apart from the commands listed above, other indispensable elements of communication with the dog are praise and reprimands, uttered at the appropriate moment, immediately after a particular behavior occurred, in line with the behaviorist approach to the mechanisms of learning and habit formation (Donaldson, 2013; Włodarczyk, 2018: 18–23). Praise must be uttered in an appropriate manner, emphasizing positive emotions and satisfaction expressed by the human toward the dog. Praising the dog is intended to reinforce behaviors expected by the human in a given situation, thereby motivating the dog to repeat them on subsequent occasions. In practice, different kinds of praise are also a way of staying in touch with the dog while walking, focusing the dog’s attention on the human handler, establishing and consolidating their rapport (Magnus, 2014: 451–452): “Show the door. Where is the door? Yes, here is the door. Great! Yes, here is the door. Super dog!”; “Good dog. Yes! Here, you get [the treat], here you are. Very nice.”; [While putting on the harness] “We are going for a walk. Come [dog’s name, diminutive]. Good doggie. Yes! Put your head in. Good. Wait, I have to strap you in. And now wait. Good doggie.”

Reprimands or verbal corrections are intended to lead to the extinction of those behaviors that are potentially or actually hazardous to the visually impaired handler and to discourage the dog from repeating them in the future. Most often, they apply to the situation when the dog fails to target key elements of space (kerbs, stairs, pedestrian crossings), pulls toward other dogs, keeps sniffing impulsively, reaches for food lying on the pavement – all in all, when the dog is not focused on tasks designated by the handler. Reprimands are uttered in a firm, strong tone, but not shouting; they express the handler’s displeasure. The simplest verbal correction is “no!”, also taking the form of “don’t!”. However, most often, dog handlers use more elaborate cues, frequently formulating them as questions: “Oi, behave!”, “Why are you stepping to the side? No. Straight. Forward.”, “What are you doing! What are you doing? Get back!”, “What are you doing? Don’t! You’re at work. What was that?”. Such messages most often summoned the dog’s attention and after momentary interest in a particular element of the environment, the dog usually moved their gaze back to the handler and resumed their designated guiding role.

I also noted the presence of negotiation with the dog as a form of verbal communication, often in conjunction with verbal correction, but usually constituting its softer, dialogic variety. Such manner of speaking to the dog often stemmed from the handler interpreting the dog’s behavior as erroneous or inaccurate, sometimes attributing specific intentions (e.g., when the dog stopped on a straight section of the route and the handler interpreted it as an attempt to ‘cheat’ and elicit a treat for targeting a non-existent obstacle) (see Magnus, 2016: 281). The negotiation mode was also used to address the dog’s needs, especially physiological: “Left. No, no. [The handler reflects for a moment]. Aha, wait, you’re going well”, “Forward, don’t pretend, there’s nothing here. Don’t cheat.”, “[Name of dog], busy busy. Don’t sniff, just pee. You did at 2 p.m. Well, where is it that are you peeing? There’s no grass here.”, “Hup-up! We keep running, hup-up! Well, your pace so not really up to it today, [dog’s name]. Oh yes, great!”

The above messages are examples of what I was able to hear during my walk-alongs with guide dogs and their handlers. They made me realise the extent to which their cooperation relies on attentiveness and communication on the part of the handlers, organically attuned to the dog’s current actions and reactions. Talking to the dog appeared to go far beyond issuing learned commands. In practice, it largely served to call the dog back to the relational order, in the role of a guide, preventing the dog from ‘stopping work’, ‘switching off’, and following their instincts. Similar procedures have also evidenced the contextualized nature of the cooperation and rapport between the visually impaired handler and their guide dog, based on learned skills and embodied habits of the dog, but requiring constant care and nurturing on the part of the human. Without these elements, a trained guide dog can quickly become unmotivated to work with the handler and start paying attention to other elements of the surrounding environment, thus falling outside the framework originally set by the trainer.

“I can feel it on the harness” – observing non-verbal communication between a human handler and the guide dog

Observing non-verbal communication between the visually impaired handler and the guide dog was an important part of our walk-alongs (Magnus, 2014; Due, 2023). I was particularly interested in ways in which the study participants sensed and interpreted the dog’s behavior – movements, changes in posture, sounds. Whenever I had an opportunity, I inquired about specific elements of their work with the dog, usually asking them to tell me what exactly they ‘read’ from a particular dog behavior and why they reacted in one way or another.

The main information transmitter when working with a guide dog is the harness handle, usually held by left hand, through which people with visual impairment sense both horizontal and vertical movements of the dog, including turns, directional changes, targeting stairs leading up, changes in pace (Curtis, 2017: 104–105; Michalko, 1999: 122). To enhance their sensibility to these cues, many participants tended to hold the handle very gently, not gripping it firmly with their fingers. For some handlers who held the leash in their right hand, the leash was also an important relay of the dog’s movements. Some of these signals were visible to me. Some, however, were invisible from the outside during the walk-along and only became apparent when I reviewed the video footage in slow motion. This included, for example, sensing the last step down the stairs, a difficult element for novice handlers, especially if the dog has not been taught to target it by stopping.

– [Researcher] Tell me please, because I’ve always been curious about this… The stairs [leading down] end and you don’t make that extra high step, you just know there’s a walkway. Because sometimes people take that king of high step at the end. How do you know that was the end of the staircase?
– Wait, I need to think about it. [Another staircase leading down]. I think [I know] from the dog.
– [Researcher] But from what exactly?
– That he’s suddenly not going down like this [hand gesture down at 45 degrees], but like this, evenly [horizontal hand gesture]. I can feel it from his posture. It’s generally about the dog position. It’s all thanks to the harness. If I didn’t have this, I would not be connected to the dog.

The handle and leash also allowed handlers to sense other changes in the dog’s walking pattern or head movements. As a rule, such signals increased the handler’s alertness, as they could signify arousal or the dog’s focus on elements such as other dogs, birds, children, balls, food lying on the ground. For some dogs, these types of reactions were unambiguous, sensed as strong tugging and pulling toward a chosen target. Other dogs reacted in more subtle ways that over time became gradually clear to their handlers.

[The dog barking at about 15 metres away] Wait, wait, there is a dog walking by. But, exceptionally, she’s not doing anything. Oh, look, now she’s swinging her shoulders more. Even, even. She’s walking so much more brashly, like she is ready to pounce on the dog. You can feel it, you can really sense it. I know her well enough already.

In this context, the study participants appreciated dogs that gave such signals well in advance by lowering their body, slowing down abruptly or characteristic squealing that only occurred when other dogs were in view. This type of behavior gave visually impaired handlers time to act and mitigate, draw the dog’s attention to the job, to bring the dog back to order. The handlers whose dogs reacted to such stimuli without warning faced a much more challenging task, which was stressful and required the handler to rapidly intervene.

Photograph 4. A visually impaired handler with a guide dog walking down a flight of stairs. The last step down is a sensitive point that requires the handler to pay more attention and sense the dog’s movements
Source: photograph by the author.

The ‘reading’ of the dog by the study participants was much easier and more effective when we traveled along their familiar routes. The handler and the dog appeared more relaxed, confident, and fluent in their movements and messages. In some cases, when following such routes, the dogs were switching into the ‘leading by heart’ mode, which meant anticipating their handler’s decisions and taking the initiative as to the direction of the walk. In such ‘leading by heart’ mode, the dogs would slow down or stop at ‘points of choice’, i.e., places where their frequented routes crossed. The dog would then glance at the handler’s face, ‘asking’ for further directions.

Observing such situations, I noticed how dogs and handlers mapped and used the affordances available in their environment and, in particular, how their decisions were influenced by the situational context, habits, or anticipation (Ingold, 1994: 13; Reed, 1994). For example, if the team was moving through an unfamiliar space and the handler uttered the command ‘door’, the dog would generally lead toward the door that was closest or most unambiguous. Following Jakob von Uexküll, we could say that the dog searched for, recognized, and acted toward those elements of the environment that, in their perception, produced an ‘entering tone’ (von Uexküll, 2010: 93–94). The handler would then scan the immediate environment or inquire if they were in the designated place (asking me during our walk-alongs, in other cases directing such questions to other passers-by). If the handler discovered that they had missed their destination, they would backtrack, set the direction, and utter the “door” command again, repeating the procedure until successful (and, when moving on their own, the handlers may also rely on the kindness of passers-by leading them to their destination).

Photograph 5. The moment when the guide dog notices another dog, gently lowers down and starts walking in a different way. Experienced guide dog handlers are able to sense such signals through the harness handle
Source: photograph by the author.

For familiar routes, I observed two scenarios. In the first one, the handler would give the command “door” and the dog, guiding ‘by heart’, would lead the handler to a particular door, such as a newsagent’s shop, even though there were other entrances to several other shops along the way. This happened in situations where the handler never or only occasionally visited other shops in a given street or shopping center, usually aiming for one specific point. Hence, with the general “door” command, the dog ignored all other entrances along the way, because, in the given context, they seemed to disappear from the dog’s attention as potential destinations.

The second scenario was used in situations where handlers used several service points that were located next to each other in an equally frequent manner. In such cases, instead of using the general command ‘door’, after which the dog would lead them to one of the familiar shops, some handlers introduced separate commands for individual destinations (e.g. ‘bakery’, ‘shop’, ‘bar’), thus avoiding extra work in finding the chosen destination. However, I should point out that these types of commands only applied to specific, familiar locations, and not to the specific type of destinations more generally. Teaching the dog to follow the command ‘pharmacy’ did not mean that the dog acquired the ability to guide the handler to any pharmacy along the route or in unfamiliar surroundings.

As the above examples demonstrate, the use of the go-along method during walk-alongs with visually impaired handlers and guide dogs made it possible for me to observe their specific ways of interacting and communicating, closely linked to particular elements of space. Also, the situations captured in this way shed new light on the stories about working with the dog that I collected through individual interviews, and, in particular, allowed me to situate them within tangible places and events.

Conclusions

In my view, the use of the go-along method in the project presented here proved to be a very appropriate choice, not only enriching the collected data with observations and audio-visual material, but also facilitating my experience – as an observer and a companion – of the intensity of the guide dog training process and the actual conditions of the dogs’ cooperation with visually impaired handlers. However, I find it difficult to imagine developing knowledge on this subject through the go-along method alone, without the use of sit-down individual interviews. These two methods proved complementary as well as they guided the research process dynamics, enriched my relationship with research participants through our shared experience, and ensured the triangulation of the collected data (Carpiano, 2009: 271; Battista, Manaugh, 2017: 17). Notably, following research participants allowed me to closely observe the dogs’ behaviors and reactions, both vis-a-vis their human handlers and other elements of their environment that, depending on the situation, constituted an obstacle, an incentive, a sign of upcoming pleasure or discomfort, an object of interest, or a cause of anxiety. Noting such behaviors supports an interpretation that guide dogs are active agents, making specific decisions depending on available resources and constraints (Horowitz, 2009; Koski, Bäcklund, 2017; Meijer, 2019). Even though my own view of the dogs’ actions has been largely guided by the narratives provided by the trainers, visually impaired handlers, or puppy raisers, I assume that this additional perspective raises an opportunity to perceive specific elements of dog-human relationships that not necessarily emerged as salient during interviews with the research participants. I also believe that such a perspective could facilitate a wider, not exclusively anthropocentric view in the overall analysis of the collected data (Weil, 2012; Lestel, 2015). It was extremely useful to complement walk-alongs with filming the study participants’ behavior on specific routes, followed by a thorough review of the collected material and efficient incorporation and linking of all the data via the MAXQDA software. Notably, filming, following, and talking to the participants was not difficult to manage and did not seem to disrupt my attention. On the contrary, the use and appropriate positioning of the small sports camera allowed me to skip note-taking on the go and made it possible to focus more on conversation and observation. I perceived the collected audiovisual material as raw data rather than a potential part of research results presentation, which made me pay little attention to the esthetic qualities of the footage or their potential evaluation.

The decision to use the audiovisual material in this way had other advantages, generating greater trust and openness on the part of the study participants, who did not need to worry about their interaction with the guide dog being evaluated by potential viewers. Admittedly, this also limited the attractiveness of the way in which I present the research outcomes, especially related to the bodily and sensory dimensions of the visually impaired handlers’ cooperation with guide dogs. In hindsight, I now understand that my ethical sensitivities – perhaps exaggerated – stemming from my reluctance to abuse my visual advantage, might have been one-sided, based solely on my own judgement, which I did not consult with the study participants. Rather than completely ruling out the dissemination of the video footage, I could have followed a more individual and contextual approach, leaving more room for choice made by the study participants themselves. I may adopt this strategy when presenting the final research outcomes by asking individual participants for permission to disseminate selected parts of footage from our walk-alongs.

The second clear limitation of the use of the go-along method related to logistics and the fact that the study participants resided in different locations in Poland and Slovakia. It was thus difficult for me to follow the daily rhythm of their lives and accompany them on their “ordinary” walks and journeys – I only managed to join 9 out of 29 dog handlers for such daily-life occasions. Hence, most of the walk-alongs were triggered in one way or another by the interviews we had scheduled. Of course, the participants were by no means ‘acting out’ in front of me and altering their cooperation with their guide dog in such presumably ‘extraordinary’, arranged conditions. The only alteration may have been the purpose and circumstances of the outing – which may have resembled a walk with a friend – without interfering with working methods, communication with the guide dog, or the traveled routes.

Just like Kusenbach (2003: 464), I am far from romanticizing the go-along method as a way to join the ‘natural’ rhythm of the lives and perceptions of research participants. Like any other research technique, the go-along also creates a specific social situation, along with a specific space for exchange and dialog, and a specific division of roles and expectations, which would not occur without the researcher’s invitation and the openness of the research participants. However, as the above examples demonstrate, walk-alongs can reveal aspects of life and insights of research participants that may not have been expressed or noticed in less dynamic circumstances. In this sense, to some extent, there was an affinity between walk-alongs themselves and what I observed during them, focusing on the interaction between visually impaired handlers and guide dogs. They all had a purpose, they required trust, attentiveness and communication, revealed additional dimensions of mobility and space, were based on both guiding and being guided, and – hopefully – were enjoyable for all parties.

Translated by Katarzyna Byłów

Acknowledgements
I wish to thank all study participants – NGO members, trainers, puppy raisers, candidates, handlers – for their openness and willingness to collaborate and to share their knowledge and experience. I am thankful to all the organizations that allowed me to accompany them during training and other activities in Poland (Fundacja na rzecz Osób Niewidomych Labrador – Pies Przewodnik, Fundacja VIS MAIOR, Fundacja Pies Przewodnik, Jerzy Przewięda Szkolenie i Tresura Psów) and in Slovakia (Psi na život, Výcviková škola pre vodiace a asistenčné psy).

Cytowanie
Kamil Pietrowiak (2024), “Follow the Dog”: Using the Go-Along Method in Research on Training and Working with Guide Dogs for People with Visual Impairment, „Przegląd Socjologii Jakościowej”, t. XX, nr 2, s. 62–89, (https://doi.org/10.18778/1733-8069.20.2.04)

^* Kamil Pietrowiak PhD, ethnographer, cultural anthropologist. He works at the Institute of Anthropology, University of Gdańsk. His research interests are related to disability studies, especially perspectives of visually impaired people, methodology, and the ethics of qualitative research. He conducts various workshops concerning senses, disability, and integration. He participates in non-academic research for cultural institutions, NGOs, and companies, e-mail:kamil.pietrowiak@ug.edu.pl

© by the author, licensee University of Lodz – Lodz University Press, Lodz, Poland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license CC-BY-NC-ND 4.0 (https://creativecommons.org/licenses/by-nc-nd/4.0/)